At 05:57 PM 11/12/00 +0100, Amara wrote:
> Evolution of Music and Myth both start with language
>
> Myths <== From Language ==> Music
> emphasizes emphasizes
> meaning sound
I.e. myth = semantics,
music = ? syntactics? (not quite),
language = ? pragmatics? (primarily)
I think while music obviously stresses sound, it's the *patterned
structures* of the sounds that are most relevant, especially in so far as
those structures derive from, reinforce, morph or otherwise play with deep
internal physiological periodicities and arhythmias. And those in turn
(since monitoring them, if only unconsciously, is a register of one kind of
important *meaning* to an organism) can embue music with [faux?]
meaningfulness at a truly profound level: hence the different impacts of
sacred or highly sexualised or meditative music.
(I liked the brief comments about vectors and wavelets. It resonated
somewhat with some stuff I wrote in a book once, which I'll glue in below
for the hell of it.)
Damien Broderick
=========================
Structuralist literary theory elaborated various binary models of the way
in which fictional sentences and larger structures are generated and
received. Despite the problems that we've discussed with binary
oppositions, one of these distinctions remains a fruitful site for
investigation. On Roman Jakobson's classical Prague-semiotics analysis,
language (and, by extension, all coded signifying systems) operates between
two poles, metaphor and metonymy. Each is a principle of verbal
substitution, the first by actual or fancied resemblance (crown standing
for dental prosthesis, because it `caps' the tooth), the second by
part-for-whole, or whole-for-part, based on contiguity or connection (the
crown standing for the king, because only a monarch may wear one). In his
description of the construction of message strings, Jakobson ([1956], 1971)
suggestively conflated metaphor with selection (where a signifier is
selected from a list or lexical `paradigm' of alternatives) and metonymy
with combination (where one word or concept bonds with another that is in
some respect adjacent in meaning-space - for example, a fragment standing
in for the whole).
In the first instance, then, both metaphor and metonymy (or synecdoche)
are figures of speech, ways of replacing one term by another. Drawing on
the Hesse net or the collective-computation models described above, we
might say that two such concepts are linked by lines of association which
bear differing probability weights, or brought together in a shared
multi-dimensional `basin of attraction'. If one enters a search at any
given point in the network, some words will be retrieved by their
structural resemblance (as previously constructed and indexed by the net),
some by partial association, some by explicit negation of resemblance (and
these, it should be noted, will be strongly linked, rather than, say, being
stored separately with very low probability of association: not-A being
salient in a different way from non-A).
If this is plausible, we see that the Hesse net location of `hat' will
forge synaptic connections with the neurons storing various other terms:
some linking by structural resemblance at the conceptual level (`bowl',
`scarf') or the phonic/lexical level (`cat', `drat'), others by
associations of use or familiarity (`head', `wife' - permitting Oliver
Sacks's patient to mistake his wife for his hat), or connections of part
and whole (`brim', `garment trade', `label'), and so on.
Within these parameters, a metaphor is a word-choice which evokes another
word by a fancied underlying structural equivalence in the signifieds or
concepts they tag, while a metonymy replaces one word with another by
virtue of some adjacency or continuity in the signifieds.
In a somewhat strained visual pun on Cartesian co-ordinates, Jakobson
created a further metaphor of his own, linking all acts of lexical
selection with a `metaphoric axis' and of sentence construction or
combination with a `metonymic axis'. The Cartesian model is apparent: a
vertical y axis, traditionally the parameter for mapping location in time,
is identified with the paradigmatic list of roughly equivalent words, while
a horizontal x axis, traditionally the parameter for graphing motion in
space, shows the syntagmatic or step-by-step activity of linking the words
chosen from the paradigmatic lexicon. The ascription is perhaps arbitrary;
a case can be made for reversing these tags, so that the time axis notates
the step-by-step chaining of syntagmata, while the space axis registers the
sites where items comprising each separate paradigm are stored, heaped
together for instant access.
Some further confusion is also evident. The activity of selecting words
from paradigmatic lists of synonyms is hardly restricted to metaphoric
substitutions but, of course, will also entail metonymic discriminations.
And while words which are combined syntagmatically in sentences thereby
acquire a measure of mutual metonymic shading, this obtains only as a
second-order effect (though a crucially important one).
Machines of the Unconscious
Under the influence of the psychoanalytic revival, a further pair of
distinctions has also been conflated into these axes: Freud's mechanisms of
condensation and displacement. The first is ventured to account for the
collapse into a single word or image (in dreams and unintentional
word-plays) of components otherwise not obviously linked. We might construe
this as the creation - or revelation - of a new paradigm containing these
several elements. The second transfers to an apparently innocuous word or
image the charged emotions properly associated with another. This is
arguably a syntagmatic procedure (Silverman, 1983, pp. 87-125). In brief:
[S]election and combination impart structure and meaning to messages
whether the repertoire of the code consists of discontinuous forms, like
DNA, or of continuous forms....
In the axis of the code selection corresponds to substitution and
similarity, the axis of metaphor.
In the axis of the message combination corresponds to contexture and
contiguity, the axis of metonymy. (Wilden, 1987b, pp. 197-8)
Citing Jakobson's analysis of the plight of aphasic patients, Wilden notes
that in `similarity defects' the procedures for combining words and
understanding their context remain effective, although synonymy is impaired
(he terms this `coding disorder'). In `contiguity defects', the reverse
occurs: patients retain access to lists of words but cannot combine them
for effective communication (a loss which Wilden terms `message disorder')
(p. 199).
On this basis, Wilden lays out a typical synoptic schema for this model:
. The axis of the code links the following functions or terms of
discrimination:
Selection and substitution, similarity, langue, simultaneity and
synchrony, paradigm, metaphor, condensation, harmony.
. The axis of the message links these alternative functions or terms of
discrimination:
Combination and contexture, contiguity, parole, succession and diachrony,
syntagm, metonymy, displacement, melody (p. 202).
The worth of these distinctions depends to some degree on the faith one has
in the models that precede their elaboration: for instance, Freud's
postulation of primary and secondary processes. (Silverman maintains that
`condensation and displacement represent the habitual response of the
primary process to similarity and contiguity, while paradigm and syntagm
constitute the normal response of the secondary process. Metaphor and
metonymy will be seen as occupying a mediate position....' (Silverman,
1983, p. 86). While for the purposes of my argument, I find the distinction
between the two classes to be of strategic value, it is worth exploring
some alternative ways to capture it.
The principal segregation is between operators of equivalence and
sequence, akin to space and time. Some alternative pairs (none of them
binary oppositions) are: list (or file), versus path (or track) (a
distinction recalling the two conjectured methods whereby, in Kosslyn's
mental imagery experiment, subjects `move' from place to place in an
imaginary landscape); map and route; frames and schemata; matrix and vector
(which might suggest an analogy with the quantum description of eigenvalues
on a state vector); position and momentum; and so perhaps, in the deepest
layers of the phallocratic imagination, uterus and penis (a typical example
of masculine appropriative - and imaginary - hierarchies).
None of these pairs is synonymous with any other, yet taken together they
suggest a strategic map on which theory might plot its tactical routes, a
list of cognitive locations from which one might choose to draw a suitable
path. Perhaps significantly, it is only now that the syntagmatic end of the
dichotomy is being granted equal analytical importance. In Beyond
Superstructuralism, for example, Richard Harland has opened an intriguing
path winding through a variety of intellectual territories, bringing them
into fruitful communication. Because his ambit is so impressively large, it
is impossible to give a fair account of his achievement - which, in turn,
is really only a ground-clearing exercise. `Most of the hard work,' he
admits candidly at the end, `remains to be done' (p. 225).
Superstructuralism, you'll recall, is a whimsical-but-serious term Harland
coined to capture the common biases and techniques of traditional
structuralism - most famously, the Lévi-Straussian variety that unmasked
binary oppositions inside myths and other cultural discourses - as well as
Theory's post-structuralisms which unpicked those oppositions in turn,
claiming that apparent polarities always enfold each other in a ceaseless
shimmer of uncertainty. Invoking a syntagmatic theory, Harland steps beyond
that story by emphasising the similarity dimension that has tended to be
overlooked in a metaphor-obsessed modernist tradition.
Theory, by and large, ruthlessly deconstructs single words, building their
several denotations into ornate narratives. Derrida's
philosophical-cum-literary career has produced a series of these fertile
vortices of ambiguity: différance, pharmakon, supplément. Like the
components of Mallarmé's symbolist poems, these words `disseminate' into an
infinite haze of reverberations. While acknowledging the merits of this
approach, Harland insists that we return our gaze to a forgotten component
in the chain: the chain itself, the sentence-string along which these words
are strung like flashing jewels: the syntagm.
The shift to a syntagmatic analysis is strikingly consequential. Meaning,
as Harland reminds us, is never found in individual words, but `stretches
across' them on a higher semantic level. An unsettling implication is that
words in a syntagm (a sentence, or chunked part of a sentence) do not add
or multiply together to create a meaningful message. Rather, each extra
word acts to subtract or narrow or refine the multitudinous connotations
that a lexicon would credit to each of the words. Far from rejoicing in the
unchecked dissemination of allusions, which Theory applauds and performs
(often with witless ingenuity), syntagmatic processes act like a series of
overlapping coloured filters, progressively cutting out one hue after
another from the spectrum, leaving at the end a crisp(ish) set of
restricted tints (or meanings) (p. 17).
Put like that, it might not seem such a startling thought. In the torrent
of turn-of-millennium thought, however, it is distinctly contrary.
Harland's impressive achievement has been to explore its implications in
fields tracking from philosophy (where it reinstates a form of
phenomenology, dispelling errors in both metaphysics and analytical
empiricism), to semantics and syntax, to literary readings, textual
interpretation and the woes of deconstruction. It is too soon, however, to
know if his unorthodox suggestion will prove acceptable to the rival
theoreticians of various superstructural schools.
Modernism, Postmodernism, Antimodernism
Metaphor and metonymy, or paradigm and syntagm, are resonant figures, then,
for more than figuration itself. Novelist and theorist David Lodge (1981,
pp. 3-16) discerns a regular pendulum swing in critical and creative
literary taste between two poles: the `transparent' or `realist', and the
`obscure', `symbolist' and inward. Since the second is `modernist', the
first, contemporary with it but never on exactly the same footing (being
sometimes higher in estimate, sometimes lower), Lodge dubs `antimodernist'.
The arrival of the `postmodernist' (which Christine Brooke-Rose [1981, p.
345], a cautious admirer of Lodge's typologies, has naughtily suggested
means `most-modernist')
continues the modernist critique of traditional realism, but it tries to
go beyond or around or underneath modernism, which for all its formal
experiment and complexity held out to the reader the promise of meaning, if
not of a meaning. (Lodge, 1981, p. 12)
What powers this pendulum? Lodge suggests that it is the resurgent
consumer (and producer) need for novelty, defamiliarisation, coupled with
the bipolar nature of the linguistic function itself. According to Lodge's
rather-too-neat story, modernism is a technique of writing and reading (and
reality-construction) founded preferentially in metaphor. Antimodernist
`realism' is given, by contrast, to figuration based on synecdoche or
metonymy. The postmodernists, predictably, push both poles to an extreme;
they `tested them, as it were, to destruction, parodied and burlesqued them
in the process of using them, and thus sought to escape from their tyranny'
(Lodge, 1981, p. 14). Arguably their escape, by definition, must be into
silence or inanity. To take a more hopeful view, escape from the
metaphor/metonymy metronome might lead into a rupture with `verities' so
deeply instilled that no-one functioning in a given episteme can readily
appreciate their arbitrariness. That escape route is commonly taken to be
post-structuralist Theory.
The Uses of Pragmatics
In the light of such investigations into categorisation and the discursive
construction of textuality (if not of reality entire), it must be asked
whether the apparently divergent impulses of referential, ideological and
deconstructive analyses can be brought to a common focus or shared account.
A persuasive synthesis of the most telling features of these disjoint
approaches has been sketched by Christopher Butler (1984). Like many of the
more interesting post-deconstructive discussions of how we read, his
synopsis is regulated by pragmatism:
For if the text has multiple implications... and if the interpreter can
also choose amongst structuralist, deconstructionist, Marxist, and liberal
moralist frameworks... then the apparent anarchy of pluralism can at least
be brought into some order, and lose its air of indifferentism, if we ask
that... interpreters should be as clear as possible about the ends for
which they interpret. (ibid., x)
Tellingly, use becomes the final arbiter. Yet `use' has application only
within a pre-established context of situation and ends. Language may be
made to turn on itself in search of fissures and aporias, but such
sceptical deconstructive play is inevitably grounded in the fact that
language is a social and socialising tool through which, in large part, we
humans are constituted as persons (`subjects'), and with which we
coordinate our physical and mental activities in the supraindividual
functions of our collectivity.
Those joint activities in turn may be construed as `texts', but it is
arguable that such texts are written in a metalanguage of higher logical
type not amenable to disruption from levels subordinate to their own. (For
instance, the apparent paradox generated by a sign which states `Disregard
This Sign' vanishes when we see that the shifter particle `this' cannot
meaningfully be applied self-reflexively; it must gesture to a referent of
lower logical type. It makes perfect sense to write `Disregard this sign on
Sundays' if the words are followed by an injunction applying to all other
days, but that is because we see that `this' then refers to the injunction
thus denoted, not to the sign in toto. With the disappearance of the
oppressive sense of paradox, we can register the sign as a witty or parodic
demolition of authoritarian injunction). If so, the best that can be done
is to make pragmatic choices, the terms of which are declared within
superordinate systems.
Butler's case begins with a preliminary stipulation, later challenged, of
`common-sense' linguistic referentiality. How do we understand? `[T]he
meaning of sentences, and above them of discourse and texts, has to be
interpreted in context, as relevant to a situation, which thereby activates
a complex of beliefs in the speaker and hearer or author and reader'
(Butler, 1984, p. 1). This complex of enabling beliefs includes logical
implicature, such that all and only those canonical implications of a given
sentence can be assumed as infolded within the sentence. Familiarity, so to
speak, breeds context, and context familiarity.
It is easy to see, although Butler does not make the point, that this
account faces several recurrent philosophical difficulties. To utter a
sentence which describes the appearance of a golden mountain or a unicorn
is to seem to insist on the same ontological status for the referents of
their signifieds as would be appropriate to observables such as ordinary
mountains or horses - an error which considerable 20th century analysis has
worked hard to elucidate and expel. In short, the fact that language
invokes logical implicature tends to set boundary conditions rather than
reveal novelties. (Perhaps this is explained by Wilden's model in which
sensory experience takes analog forms which parse by `more-or-less' while
communication takes discontinuous digitalised forms, analytic logic being
nothing if not digital.)
Beyond the strict constraints of logic we encounter pragmatic `contextual
determinants of implicit meaning' (Butler, 1984, p. 1). No intelligible
text-fragment is an island. Each is uttered in a language whose grammar and
lexicon are shared by other speaker/writer/readers, for even the severest
idiolect is a transformed version of a common tongue:
Thus our model for interpretation must include not only a body of
linguistic knowledge (a grammar), but also a background of knowledge and
belief (an `encyclopaedia' for short) which can if necessary be made
explicit in interpretation. (p. 2)
The informational voids and gaps in any textual syntagmatic string are
enormous by comparison with the richness of information we construe (not
exactly the same as `construct') from the material provided. The coding
involved is always a sort of metonymic condensation. Consider a homely
example: the word `dog' is hugely overdetermined for both addresser and
addressee, and in quite different ways for each. Possibly it includes every
dog each has ever encountered in life, in pictorial illustration, in
fiction. Each bears its freight of affect and further conceptual
associations, adding to a vast ramifying synaptic and semiotic root-system
(or computational-energy `landscape'). These arrays, stored in individual
heads, just sufficiently resemble those in other heads that generally a
salient response can be evoked, it is hoped, without a great deal of
backing and filling. Plainly, though, the word `dog' alone cannot provoke a
particularly sharp determinate response or `concretisation'. In a fiction
text, much of the mood and set shaped by preliminary and subsequent codings
help govern the way a reader actualises this schematic element. But each of
those in turn also depends, like the struts of a geodesic dome, on the
tensility of its neighbours.
Categories and Prototypes
Eleanor Rosch's important researches into categorisation bear on this
issue. Gardner summarises her findings thus:
Categories are built around a central member or prototype - a
representative example of that class which shares the most features with
other members of the category while sharing few, in any, features with
elements drawn from outside the class. (Howard Gardner, 1985, p. 346)
Rosch discerns several moments in this process. There is a basic level that
is defined by the ease with which children and newcomers acquire its grasp:
`dog' is a primitive on this level. Above this is a superordinate level,
such as `animal', `mammal', `pet', `food' (or `inedible'), or perhaps
`totem'. Thus, as Barry Barnes has shown in his remarks on Hesse nets, a
cassowary may be assimilated to the prototype of a totemic group which has
no other avian members, while a bat may be drawn into the group of `winged
animals' that we would not parse as `birds'.
Below the basic level is a subordinate level that more closely specifies
the individual case, such as `German Pointer', `Pointer-Heeler Cross',
`Imme', `Rommy' or `Meggy', and so on. However, as we have noted
repeatedly, `categories in the real world tend to have fuzzy boundaries and
to blend into one another' (ibid.).
For all that, these categories are not, as a Whorfian or Saussurean would
demand, entirely a matter of free convention. Rosch's work on colour
discrimination has demonstrated that with visual perception `what looks
similar [both within a culture, and between cultures] is not in the least
arbitrary. For example, our visual system is so designed as to treat
certain reds as better than others, and to draw the line between red and
orange [etc] at specific points on what looks (to the instruments of the
physical scientist) like a continuous spectrum' (p. 347). In other words,
the human perceptual system is hard-wired to transduce certain of Wilden's
smoothly varying analog information sources into crisply separate digital
discriminations, or to try its damnedest to do so.
In any event, the task of interpreting a piece of fiction clearly shares
certain characteristics with the task of interpreting any other organised
sensory input. The internal organisation of a text constrains each act of
parsing we perform, rendering some later `actualisations' more probable
than others, yet also able to
force a re-assessment of what we have already settled as our provisional
reading. Butler turns to the social linguist Michael Halliday for his
notion of an intratextual `co-text':
For the text is typically `internally cohesive, and it functions as a
whole as the relevant environment for the operation of the theme and
informational systems'.... I shall... attempt to distinguish between
co-text in this sense and `context', by which I shall mean the situation of
the text in the world. This will include such matters as the text's
referring `beyond itself' to `external reality'. (Butler, 1984, p. 2)
How, then, do we read a set of consecutive sentences? If they are
presented to us as a single text, we bring to bear (even if we are
card-carrying postmodernists) the preliminary assumption that they tell a
unified story - that in some sense they `hang together', whatever exotic or
polysemy-besotted sense of `unified' we might be employing. In Rumelhart's
terms, `schemata are activated, evaluated, refined or discarded'; Butler
adds: `various reader's interpretations of [such schemata's] relevance to
one another are remarkably consistent' (p. 5, citing D. Rumelhart, 1979).
His elaborated model draws on Van Dijk, de Beaugrande, Eco and others, and
distinguishes between `frame', which contains the knowledge of the
background elements that we bring to a text, and `schema', the ordered use
of these elements.
Similar postulates, as I've noted previously, have been forced upon
analysts working in artificial intelligence research, who have found that
computerised simulation systems require `common sense' before they can
perceive and act with any vestige of intelligence. One favoured method of
simulating common sense is the insertion of `scripts' - `the canonical set
of events one can expect in an often encountered setting' - which are
precisely frames of reference and `tacit knowledge' necessary to organise
`experience' into parsable chunks calling for determinable responses. What
is more, the necessity for such semantic exfoliations from each central
lexical node has been one of the principal objections raised to Chomsky's
early syntactic linguistics, and the motive for his current virtual
abolition in transformational grammar of a metastasising number of deep
structure moves in favour of semantically-loaded generative lexical
elements governed by pragmatic constraints.
Underlying this model is a neurosemantic hypothesis: that within our
brains we map the world in chunks and micro-scripts, which are
transformable through conventional coding into public texts, these in turn
receivable by brains similarly supplied with cognitive maps not too grossly
at variance with those which encoded the texts. [etc etc]
This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:50:35 MDT